1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
| %MACRO BNelson(DATA=,OUT=,VAR=,DATE=,MAXPQ=,PAR=,QMA=,METHOD=,MAXITER=,
MCOEF=,MINMOD=,PRINT=,GRAPH=,CRAW=,CTREND=,CCYCLE=,DEBUG= );
/************************************************************************/
/* BNELSON: Cycle extraction with Beveridge-Nelson method. */
/************************************************************************/
/* This macro implements the Beveridge-Nelson decomposition model of a */
/* time series into Trend and Cycle. */
/* The basic output is a SAS dataset with the series and its components.*/
/* On request, you obtain some graphs. */
/*----------------------------------------------------------------------*/
/* Author: Dominique LADIRAY (July 2001) */
/************************************************************************/
/* MACRO PARAMETERS */
/************************************************************************/
/* */
/* DATA = SAS dataset containing the input time series. By default */
/* the _LAST_ dataset is used. */
/* OUT = Output SAS dataset with the series and its components. By */
/* default OUT=_result_. The components are named BN_Trendn */
/* for the trend and BN_Cyclen for the cycle where n refers */
/* to the ajusted ARIMA (p,1,q) model. In case several models */
/* are possible, a decomposition is done for each model (see */
/* MAXPQ). */
/* VAR = The time series to be decomposed. It must be numeric. */
/* DATE = This variable is used for the graphs and for the Ljung-Box */
/* test for residuals. Therefore, it must be precised and */
/* must be a SAS date variable (this point cannot be checked).*/
/* MAXPQ = You can ask for an automatic ajustment of the (p,1,q) */
/* ARIMA model. The orders p and q will be determined with */
/* the ESACF and SCAN method available in the ESTIMATE */
/* statement of PROC ARIMA. MAXPQ precises the maximum value */
/* of the orders p and q. If the parameters PAR or QMA and */
/* MAXPQ are not specified, then an automatic modelling is */
/* performed with MAXPQ=10. */
/* PAR = In case you want to specify the (p,1,q) ARIMA model to be */
/* ajusted, the parameter PAR= precises the order of the AR */
/* part. PAR must be positive. */
/* QMA = In case you want to specify the (p,1,q) ARIMA model to be */
/* ajusted, the parameter QMA= precises the order of the MA */
/* part. */
/* PAR and QMA must be precised together otherwise an */
/* automatic modelling is performed (see MAXPQ) */
/* METHOD = The method to be used in the ARIMA estimation. Possible */
/* values are CLS (the default), ULS and ML. */
/* MAXITER = Maximum number of iteration in the estimation. By default */
/* MAXITER=&maxiter. */
/* MCOEF = The adjusted ARIMA model must be converted in a MA model. */
/* MCOEF is the number of coefficients you require for the */
/* MA model. By default MCOEF=100. */
/* MINMOD = The adjusted ARIMA model must be invertible. The roots of */
/* the AR polynomial should be outside the unit circle. */
/* MINMOD is the minimum value required for the modulus of */
/* the roots. By default, MINMOD=0.98. */
/* PRINT = If YES, you get some printed output from the automatic */
/* ARIMA modeling. By default PRINT=yes. */
/* GRAPH = If YES, you get some graphs of the components. By default, */
/* GRAPH=yes. */
/* CRAW = Color of the raw series. Default BLACK. */
/* CTREND = Color of the Trend. Default RED. */
/* CCYCLE = Color of the cycle. Default BLUE. */
/* For these last three options, use a valid SAS/Graph color. */
/* DEBUG = Debugging option */
/* */
/************************************************************************/
/* REFERENCES : */
/* Beveridge S., Nelson C.R. (1981) */
/* A New Approach to Decomposition of Economic Time Series into */
/* Permanent and Transitory Components with Particular Attention to */
/* Measurement of the 'Business Cycle'. */
/* Journal of Monetary Economics, 7, pages 151-174. */
/************************************************************************/
%IF (&debug NE) %THEN %DO;OPTIONS MPRINT NOTES;%END;
/************************************************************************/
/* SASNOM Macro, to check if a name is a valid SAS name */
/************************************************************************/
%MACRO sasnom(_IN_=,TYPE=);
%GLOBAL _ret_;
%LET _ret_=0;
%IF &SYSVER < 7 %THEN %LET max=8;%ELSE %LET max=32;
%LET TABLE=%UPCASE(%TRIM(&_in_));
%IF (%UPCASE(&type) EQ DATASET) %THEN %DO;
%LET INDEX=%INDEX(&TABLE,.);
%IF (&INDEX GT 1) %THEN %DO;
%LET lib=%SUBSTR(&TABLE,1,&index-1);
%LET TABLE=%SUBSTR(&TABLE,&INDEX+1);
%LET c1=%LENGTH(&lib);
%LET c2=%VERIFY(%SUBSTR(&lib,1,1),'_ABCDEFGHIJKLMNOPQRSTUVWXYZ');
%LET c3=%VERIFY(&lib,'_0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ');
%IF ((&c1 > &max) OR (&c2 > 0) OR (&c3 > 0)) %THEN %LET _ret_=1;
%END;
%END;
%LET c1=%LENGTH(&TABLE);
%LET c2=%VERIFY(%SUBSTR(&TABLE,1,1),'_ABCDEFGHIJKLMNOPQRSTUVWXYZ');
%LET c3=%VERIFY(&TABLE,'_0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ');
%IF ((&c1 > &max) OR (&c2 > 0) OR (&c3 > 0)) %THEN %LET _ret_=1;
%MEND;
/************************************************************************/
/* Reading and checking of parameters */
/************************************************************************/
/* Checking parameter data */
/************************************************************************/
%IF (%LENGTH(&DATA ) EQ 0) %THEN %LET DATA = _last_ ;
%ELSE %DO;
%IF (%LENGTH(%QSCAN(%BQUOTE(&DATA ),2,%STR( ))) NE 0) %THEN %DO;
%PUT ERROR: You must precise just one value FOR the DATA (&DATA ) parameter;
%GOTO finmacro;
%END;
DATA _NULL_;
SET &DATA (OBS=1);
RUN;
%IF (&syserr NE 0) %THEN %DO;
%PUT ERROR: There IS a problem IN the DATA (&DATA ) parameter;
%PUT ERROR: Non valid SAS dataset? Non valid SAS name? ;
%GOTO finmacro;
%END;
%END;
/* Checking parameter out */
/************************************************************************/
%IF (%LENGTH(&out ) EQ 0) %THEN %LET out = _result_ ;
%ELSE %DO;
%IF (%LENGTH(%QSCAN(%BQUOTE(&out ),2,%STR( ))) NE 0) %THEN %DO;
%PUT ERROR: You must precise just one value FOR the out (&out ) parameter;
%GOTO finmacro;
%END;
%sasnom(_IN_=&out ,TYPE=DATASET);
%IF (&_ret_ NE 0) %THEN %DO;
%PUT ERROR: The out (&out ) parameter must be a valid SAS dataset name;
%GOTO finmacro;
%END;
%END;
/* Checking parameter var */
/************************************************************************/
%IF (%LENGTH(&var ) EQ 0) %THEN %DO;
%PUT ERROR: The var parameter must be precised;
%GOTO finmacro;
%END;
%ELSE %DO;
%IF (%LENGTH(%QSCAN(%BQUOTE(&var ),2,%STR( ))) NE 0) %THEN %DO;
%PUT ERROR: You must precise just one value FOR the var (&var ) parameter;
%GOTO finmacro;
%END;
DATA _NULL_;
SET &DATA (KEEP=&var OBS=1);
ARRAY _xx_(*) &var ;
RUN;
%IF (&syserr NE 0) %THEN %DO;
%PUT ERROR: There IS a problem IN the var (&var ) parameter;
%PUT ERROR: Non valid SAS dataset? Non valid SAS name? Wrong variable type? Non valid SAS list?;
%GOTO finmacro;
%END;
%END;
/* Checking parameter date */
/************************************************************************/
%IF (%LENGTH(&date ) EQ 0) %THEN %DO;
%PUT ERROR: The date parameter must be precised;
%GOTO finmacro;
%END;
%ELSE %DO;
%IF (%LENGTH(%QSCAN(%BQUOTE(&date ),2,%STR( ))) NE 0) %THEN %DO;
%PUT ERROR: You must precise just one value FOR the date (&date ) parameter;
%GOTO finmacro;
%END;
DATA _NULL_;
SET &DATA (KEEP=&date OBS=1);
ARRAY _xx_(*) &date ;
RUN;
%IF (&syserr NE 0) %THEN %DO;
%PUT ERROR: There IS a problem IN the date (&date ) parameter;
%PUT ERROR: Non valid SAS dataset? Non valid SAS name? Wrong variable type? Non valid SAS list?;
%GOTO finmacro;
%END;
%END;
/* Checking parameter maxpq */
/************************************************************************/
%IF (%LENGTH(&maxpq ) EQ 0) %THEN %LET maxpq = 10 ;
%ELSE %DO;
%IF (%LENGTH(%QSCAN(%BQUOTE(&maxpq ),2,%STR( ))) NE 0) %THEN %DO;
%PUT ERROR: You must precise just one value FOR the maxpq (&maxpq ) parameter;
%GOTO finmacro;
%END;
%IF (%DATATYP(&maxpq ) EQ CHAR) %THEN %DO;
%PUT ERROR: The maxpq (&maxpq ) parameter must be numeric;
%GOTO finmacro;
%END;
%ELSE %IF (&maxpq LE 0 ) %THEN %DO;
%PUT ERROR: You must have &maxpq > 0 ;
%GOTO finmacro;
%END;
%END;
/* Checking parameter mcoef */
/************************************************************************/
%IF (%LENGTH(&mcoef ) EQ 0) %THEN %LET mcoef = 100 ;
%ELSE %DO;
%IF (%LENGTH(%QSCAN(%BQUOTE(&mcoef ),2,%STR( ))) NE 0) %THEN %DO;
%PUT ERROR: You must precise just one value FOR the mcoef (&mcoef ) parameter;
%GOTO finmacro;
%END;
%IF (%DATATYP(&mcoef ) EQ CHAR) %THEN %DO;
%PUT ERROR: The mcoef (&mcoef ) parameter must be numeric;
%GOTO finmacro;
%END;
%ELSE %IF (&mcoef LE 0 ) %THEN %DO;
%PUT ERROR: You must have mcoef > 0 ;
%GOTO finmacro;
%END;
%END;
/* Checking parameter par */
/************************************************************************/
%IF (%LENGTH(&par ) NE 0) %THEN %DO;
%IF (%LENGTH(%QSCAN(%BQUOTE(&par ),2,%STR( ))) NE 0) %THEN %DO;
%PUT ERROR: You must precise just one value FOR the par (&par ) parameter;
%GOTO finmacro;
%END;
%IF (%DATATYP(&par ) EQ CHAR) %THEN %DO;
%PUT ERROR: The par (&par ) parameter must be numeric;
%GOTO finmacro;
%END;
%ELSE %IF (&par LT 0 ) %THEN %DO;
%PUT ERROR: You must have par >= 0 ;
%GOTO finmacro;
%END;
%END;
/* Checking parameter qma */
/************************************************************************/
%IF (%LENGTH(&qma) NE 0) %THEN %DO;
%IF (%LENGTH(%QSCAN(%BQUOTE(&qma ),2,%STR( ))) NE 0) %THEN %DO;
%PUT ERROR: You must precise just one value FOR the qma (&qma ) parameter;
%GOTO finmacro;
%END;
%IF (%DATATYP(&qma ) EQ CHAR) %THEN %DO;
%PUT ERROR: The qma (&qma ) parameter must be numeric;
%GOTO finmacro;
%END;
%ELSE %IF (&qma LT 0 ) %THEN %DO;
%PUT ERROR: You must have qma >= 0 ;
%GOTO finmacro;
%END;
%END;
/* Checking parameter method */
/************************************************************************/
%IF (%LENGTH(&method ) EQ 0) %THEN %LET method = CLS;
%ELSE %DO;
%LET method=%UPCASE(&method);
%IF (%LENGTH(%QSCAN(%BQUOTE(&method ),2,%STR( ))) NE 0) %THEN %DO;
%PUT ERROR: You must precise just one value FOR the method (&method ) parameter;
%GOTO finmacro;
%END;
%IF (%DATATYP(&method ) EQ NUM) %THEN %DO;
%PUT ERROR: The method (&method ) parameter must be character;
%GOTO finmacro;
%END;
%ELSE %IF ((&method NE CLS) AND (&method NE ULS) AND (&method NE ML)) %THEN %DO;
%PUT ERROR: The method (&method ) parameter must be choosen among (CLS ULS ML);
%GOTO finmacro;
%END;
%END;
/* Checking parameter maxiter */
/************************************************************************/
%IF (%LENGTH(&maxiter ) EQ 0) %THEN %LET maxiter = 100 ;
%ELSE %DO;
%IF (%LENGTH(%QSCAN(%BQUOTE(&maxiter ),2,%STR( ))) NE 0) %THEN %DO;
%PUT ERROR: You must precise just one value FOR the maxiter (&maxiter ) parameter;
%GOTO finmacro;
%END;
%IF (%DATATYP(&maxiter ) EQ CHAR) %THEN %DO;
%PUT ERROR: The maxiter (&maxiter ) parameter must be numeric;
%GOTO finmacro;
%END;
%ELSE %IF (&maxiter LE 0 ) %THEN %DO;
%PUT ERROR: You must have maxiter > 0 ;
%GOTO finmacro;
%END;
%END;
/* Checking parameter print */
/************************************************************************/
%IF (%LENGTH(&print ) EQ 0) %THEN %LET print = YES ;
%ELSE %DO;
%LET print=%UPCASE(&print);
%IF (%LENGTH(%QSCAN(%BQUOTE(&print ),2,%STR( ))) NE 0) %THEN %DO;
%PUT ERROR: You must precise just one value FOR the print (&print) parameter;
%GOTO finmacro;
%END;
%IF (%DATATYP(&print ) EQ NUM) %THEN %DO;
%PUT ERROR: The print (&print) parameter must be character;
%GOTO finmacro;
%END;
%ELSE %IF ((&print NE YES) AND (&print NE NO)) %THEN %DO;
%PUT ERROR: The print (&print) parameter must be choosen among (YES NO);
%GOTO finmacro;
%END;
%END;
/* Checking parameter graph */
/************************************************************************/
%IF (%LENGTH(&graph ) EQ 0) %THEN %LET graph = YES ;
%ELSE %DO;
%LET graph=%UPCASE(&graph);
%IF (%LENGTH(%QSCAN(%BQUOTE(&graph ),2,%STR( ))) NE 0) %THEN %DO;
%PUT ERROR: You must precise just one value FOR the graph (&graph) parameter;
%GOTO finmacro;
%END;
%IF (%DATATYP(&graph ) EQ NUM) %THEN %DO;
%PUT ERROR: The graph (&graph) parameter must be character;
%GOTO finmacro;
%END;
%ELSE %IF ((&graph NE YES) AND (&graph NE NO)) %THEN %DO;
%PUT ERROR: The graph (&graph) parameter must be choosen among (YES NO);
%GOTO finmacro;
%END;
%END;
/* Checking parameter craw */
/************************************************************************/
%IF (%LENGTH(&craw ) EQ 0) %THEN %LET craw = black ;
%ELSE %DO;
%IF (%LENGTH(%QSCAN(%BQUOTE(&craw ),2,%STR( ))) NE 0) %THEN %DO;
%PUT ERROR: You must precise just one value FOR the craw (&craw ) parameter;
%GOTO finmacro;
%END;
%IF (%DATATYP(&craw ) EQ NUM) %THEN %DO;
%PUT ERROR: The craw (&craw ) parameter must be character;
%GOTO finmacro;
%END;
%END;
/* Checking parameter ctrend */
/************************************************************************/
%IF (%LENGTH(&ctrend ) EQ 0) %THEN %LET ctrend = red ;
%ELSE %DO;
%IF (%LENGTH(%QSCAN(%BQUOTE(&ctrend ),2,%STR( ))) NE 0) %THEN %DO;
%PUT ERROR: You must precise just one value FOR the ctrend (&ctrend ) parameter;
%GOTO finmacro;
%END;
%IF (%DATATYP(&ctrend ) EQ NUM) %THEN %DO;
%PUT ERROR: The ctrend (&ctrend ) parameter must be character;
%GOTO finmacro;
%END;
%END;
/* Checking parameter ccycle */
/************************************************************************/
%IF (%LENGTH(&ccycle ) EQ 0) %THEN %LET ccycle = blue ;
%ELSE %DO;
%IF (%LENGTH(%QSCAN(%BQUOTE(&ccycle ),2,%STR( ))) NE 0) %THEN %DO;
%PUT ERROR: You must precise just one value FOR the ccycle (&ccycle ) parameter;
%GOTO finmacro;
%END;
%IF (%DATATYP(&ccycle ) EQ NUM) %THEN %DO;
%PUT ERROR: The ccycle (&ccycle ) parameter must be character;
%GOTO finmacro;
%END;
%END;
/************************************************************************/
/* Some other controls */
/************************************************************************/
%LET erreur=0;
DATA _NULL_;
IF (INT(&mcoef ) ^= &mcoef ) THEN CALL SYMPUT('erreur','1');
IF (INT(&maxpq ) ^= &maxpq ) THEN CALL SYMPUT('erreur','1');
IF (INT(&maxiter ) ^= &maxiter ) THEN CALL SYMPUT('erreur','1');
%IF (&par NE) %THEN %DO;
IF (INT(&par ) ^= &par ) THEN CALL SYMPUT('erreur','1');
%END;
%IF (&qma NE) %THEN %DO;
IF (INT(&qma ) ^= &qma ) THEN CALL SYMPUT('erreur','1');
%END;
RUN;
%IF (&erreur EQ 1) %THEN %DO;
%PUT ERROR: the parameters mcoef, par, qma AND maxpq must be integer.
%GOTO finmacro;
%END;
/************************************************************************/
/* Controls end, Macro begins */
/************************************************************************/
%IF (%UPCASE(&print) NE YES)
%THEN %LET print=noprint;
%ELSE %LET print=;
/* We check for missing values and compute the series periodicity. */
/************************************************************************/
PROC FORMAT;
VALUE njm
25-35=12 50-70=6 80-100=4 115-135=3 175-195=2;
RUN;
%LET _miss_=0;
DATA _trav_;
SET &DATA END=fin;
KEEP &date &var;
RETAIN dtmin dtmax;
IF (&var ^=.) THEN DO;
ndon+1;
dtmax=&date;
IF (ndon=1) THEN dtmin=&date;
END;
IF ((&var ^= .) AND (LAG(&var) =.) AND (ndon > 1)) THEN CALL SYMPUT('_miss_','1');
IF fin THEN DO;
period=PUT((dtmax-dtmin)/ndon,njm.);
CALL SYMPUT('period',period);
END;
IF (&var ^=.);
RUN;
%IF (&period EQ ) %THEN %DO;
%PUT ERROR: Problem WITH the automatic determination of the series periodicity. CHECK your SAS date variable;
%GOTO finmacro;
%END;
%IF (&_miss_ EQ 1) %THEN %DO;
%PUT ERROR: Missing VALUES IN series &var;
%GOTO finmacro;
%END;
/* Non automatic ARIMA modeling of the series. */
/************************************************************************/
%IF ((&par NE) AND (&qma NE)) %THEN %DO;
DATA _bic_;
Model=1;
ar=∥
ma=&qma;
RUN;
%LET _nmod=1;
%LET _maxar=∥
%LET _maxma=&qma;
%LET _ar1=∥
%LET _ma1=&qma;
PROC ARIMA DATA=_trav_;
ODS SELECT NONE;
ODS OUTPUT ChiSqAuto=ChiSq ParameterEstimates=ParEst;
IDENTIFY VAR=&var(1);
ESTIMATE P=&par Q=&qma METHOD=&method MAXITER=&maxiter;
FORECAST ID=&date OUT=_arima_(KEEP=&date &var residual RENAME=(residual=Resid1)) LEAD=0;
QUIT;
ODS SELECT ALL;
%END;
/* Automatic ARIMA modeling of the series. */
/************************************************************************/
%ELSE %DO;
PROC ARIMA DATA=_trav_;
ODS SELECT NONE;
ODS OUTPUT TentativeOrders=_bic_ ;
IDENTIFY VAR=&var(1) P=&maxpq Q=&maxpq ESACF SCAN;
RUN;
QUIT;
ODS SELECT ALL;
%LET _ar=0;
%LET _ma=0;
DATA _bic_;
SET _bic_(KEEP=scan_AR scan_MA RENAME=(scan_AR=ar scan_MA=ma))
_bic_(KEEP=esacf_AR esacf_MA RENAME=(esacf_AR=ar esacf_MA=ma));
KEEP ar ma sum;
IF ar > 0;
sum=ar+ma;
RUN;
PROC SORT DATA=_bic_ OUT=_bic_ NODUP;
BY sum ma ar;
RUN;
%LET _nmod=0;
DATA _bic_;
LENGTH Model 8;
RETAIN maxar maxma 0;
SET _bic_;
KEEP Model ar ma;
model+1;
maxar=MAX(ar,maxar);
maxma=MAX(ma,maxma);
CALL SYMPUT('_ar'||LEFT(PUT(model,3.)),LEFT(PUT(ar,2.)));
CALL SYMPUT('_ma'||LEFT(PUT(model,3.)),LEFT(PUT(ma,2.)));
CALL SYMPUT('_nmod',LEFT(PUT(model,3.)));
CALL SYMPUT('_maxar',LEFT(PUT(maxar,3.)));
CALL SYMPUT('_maxma',LEFT(PUT(maxma,3.)));
RUN;
%IF (&_nmod EQ 0) %THEN %DO;
%PUT ERROR: No automatic model has been proposed BY ESACF AND SCAN options;
%GOTO finmacro;
%END;
PROC ARIMA DATA=_trav_;
ODS SELECT NONE;
ODS OUTPUT ChiSqAuto=ChiSq ParameterEstimates=ParEst;
IDENTIFY VAR=&var(1);
%DO i=1 %TO &_nmod;
ESTIMATE P=&&_ar&i Q=&&_ma&i METHOD=&method MAXITER=&maxiter;
FORECAST ID=&date OUT=for&i(KEEP=&date &var residual) LEAD=0;
%END;
RUN;
QUIT;
ODS SELECT ALL;
DATA _arima_;
MERGE
%DO k=1 %TO &_nmod;
for&k(RENAME=(residual=Resid&k))
%END;
;
BY &date;
LABEL
%DO k=1 %TO &_nmod;
resid&k=' ' forecast&k=' '
%END;
;
RUN;
%END;
/* We check now the validity of the ARIMA models. */
/************************************************************************/
DATA ParEst;
SET ParEst;
RETAIN model 0;
IF Parameter=:'MU' THEN model=model+1;
Parameter=SUBSTR(Parameter,1,2)||LEFT(PUT(Lag,3.));
RUN;
PROC TRANSPOSE DATA=ParEst(KEEP=Estimate Parameter Model) OUT=ParEst;
BY model;
ID Parameter;
RUN;
DATA ChiSq;
LENGTH Lag $ 8;
SET ChiSq;
KEEP Lag ProbChiSq Model ToLags;
RETAIN Model -1 n;
IF (ToLags=6) THEN DO;
model=model+1;
n=0;
END;
LAG='PLjung';
IF model > 0;
IF df;
IF (((&period=12) AND (ToLags=12)) OR ((&period^=12) AND (ToLags=6))
OR ((n=0) AND (ToLags > &period)))
THEN DO;
OUTPUT;
n=1;
END;
RUN;
PROC TRANSPOSE DATA=ChiSq OUT=ChiSq;
BY model;
ID Lag;
COPY ToLags;
RUN;
DATA ChiSq;
SET ChiSq;
LjungBox= (PLjung > 0.05);
RUN;
%LET _selmod=0;
DATA _bic_;
RETAIN n 0;
LENGTH Model 8 Arima $ 10;
MERGE _bic_ ParEst ChiSq;
DROP _NAME_ _LABEL_ n;
BY Model;
LABEL ar=' ' ma=' ';
Arima=COMPRESS('('||PUT(ar,3.)||',1,'||PUT(ma,3.)||')');
IF (LjungBox) THEN DO;
IF (n=0) THEN CALL SYMPUT('_selmod',LEFT(PUT(Model,2.)));
n=n+1;
selmod='*';
END;
RUN;
%IF (&print EQ) %THEN %DO;
DATA _bic1_;
LENGTH Arima $ 10 Selmod $1 LjungBox PLjung ToLags MU AR1-AR&_maxar MA1-MA&_maxma 8;
SET _bic_(RENAME=(MU0=MU));
ARRAY num{*} _NUMERIC_;
DROP i model ar ma;
DO i= 1 TO DIM(num);
num(i)=ROUND(num(i),0.001);
END;
RUN;
PROC TRANSPOSE DATA=_bic1_ OUT=_bic1_(DROP=_LABEL_) PREFIX=Model;
VAR _all_;
RUN;
DATA _bic1_;
SET _bic1_;
DROP i;
ARRAY model{*} $ model1-model&_nmod;
DO i= 1 TO DIM(model);
model(i)=RIGHT(model(i));
END;
LABEL _name_=' ';
RENAME _name_=Description;
RUN;
PROC PRINT DATA=_bic1_;
ID Description;
TITLE "Possible ARIMA models";
RUN;
TITLE;
%END;
%IF (&_selmod EQ 0) %THEN %DO;
%PUT ERROR: No automatic model has been selected;
%GOTO finmacro;
%END;
/* We compute Trend and Cycle estimates with PROC IML. */
/************************************************************************/
PROC IML;
USE _bic_;
READ ALL VAR{%DO i=1 %TO &_maxar; ar&i %END;} INTO ar;
READ ALL VAR{%DO i=1 %TO &_maxma; ma&i %END;} INTO ma;
READ ALL VAR{selmod} INTO selmod;
CLOSE _bic_;
USE _arima_;
READ ALL VAR{&var} INTO y;
READ ALL VAR{&date} INTO date;
READ ALL VAR{%DO i=1 %TO &_nmod; Resid&i %END;} INTO r;
CLOSE _arima_;
model=LOC(selmod='*');
nbmod=NCOL(model);
noms=CONCAT(J(1,nbmod,'BN_Trend'),CHAR(model,4));
noms=noms || CONCAT(J(1,nbmod,'BN_Cycle'),CHAR(model,4));
noms=COMPRESS(noms);
ar=J(nbmod,1,1)||(-ar)[model,];
ma=J(nbmod,1,1)||(-ma)[model,];
r=r[,model];
nobs=NROW(r);
npsi=&mcoef;
psi=J(npsi+1,nbmod,.);
cusum=J(npsi,nbmod,.);
DO i=1 TO nbmod;
cc=LOC(ar[i,]^=.);
dd=LOC(ma[i,]^=.);
psi[,i]=T(RATIO(ar[i,cc],ma[i,dd],npsi+1));
psi2=psi[2:npsi+1,i];
cusum[,i]=CUSUM(psi2[npsi:1])[npsi:1];
END;
cycle=J(nobs,nbmod,.);
maxobs=MIN(npsi,nobs);
DO i= nobs TO 2 BY -1;
min=MIN(i,maxobs);
rsub=r[i:MAX(i-min+1,2),];
csub=cusum[1:NROW(rsub),];
cycle[i,]=(rsub # csub)[+,];
END;
result=date || y || (REPEAT(y,1,nbmod)-cycle) || cycle;
names={"&date" "&var"} || noms;
CREATE &out FROM result[COLNAME=names];
APPEND FROM result;
CLOSE &out;
QUIT;
/* Graphs on request. */
/************************************************************************/
%IF (%UPCASE(&graph) EQ YES) %THEN %DO;
%IF %SYSFUNC(CEXIST(&var)) %THEN %DO;
PROC CATALOG C=&var KILL;
QUIT;
%END;
DATA _NULL_;
SET _bic_(WHERE=(selmod='*'));
CALL SYMPUT('model'||LEFT(PUT(_N_,3.)),arima);
CALL SYMPUT('trend'||LEFT(PUT(_N_,3.)),'BN_Trend'||LEFT(PUT(model,3.)));
CALL SYMPUT('cycle'||LEFT(PUT(_N_,3.)),'BN_Cycle'||LEFT(PUT(model,3.)));
CALL SYMPUT('nmod',LEFT(PUT(_N_,3.)));
RUN;
GOPTIONS CBACK=white BORDER noDISPLAY DEVICE=win FTEXT=swissb noPROMPT;
SYMBOL1 I=JOIN COLOR=&craw WIDTH=1 VALUE=none ;
SYMBOL2 I=JOIN COLOR=&ctrend WIDTH=1 VALUE=none ;
SYMBOL3 I=JOIN COLOR=&ccycle WIDTH=1 VALUE=none ;
AXIS1 LABEL=NONE;
PROC GPLOT DATA=&out GOUT=&var UNIFORM;
%DO i=1 %TO &nmod;
PLOT &var*&date &&trend&i*&date /FRAME NAME="orig_s&i" VAXIS=axis1 OVERLAY;
FORMAT &date MONYY5.;
TITLE "Variable &var and Trend, model &&model&i";
RUN;
PLOT &&cycle&i*&date /FRAME NAME="cycle&i" VAXIS=axis1;
FORMAT &date MONYY5.;
TITLE "Cycle of variable &var, model &&model&i";
RUN;
%END;
QUIT;
GOPTIONS display;
PROC GREPLAY IGOUT=&var TC=&var NOFS GOUT=&var;
TDEF newtwo DES='two squares of equal size'
1/LLX=1 LLY=50 ULX=1 ULY=99 URX=99 URY=100 LRX=99 LRY=50 color=grey
2/LLX=1 LLY=1 ULX=1 ULY=50 URX=99 URY=50 LRX=99 LRY=1 color=grey
;
TEMPLATE newtwo;
%DO i=1 %TO &nmod;
TREPLAY 1:orig_s&i 2:cycle&i;
RUN;
%END;
QUIT;
PROC GPLOT DATA= &out GOUT=&var UNIFORM;
%DO i=1 %TO &nmod;
PLOT &var*&date &&trend&i*&date /FRAME NAME="both&i" VAXIS=axis1 OVERLAY;
PLOT2 &&cycle&i*&date /FRAME NAME="double&i" VAXIS=axis1;
FORMAT &date MONYY5.;
TITLE "Variable &var, Estimated Trend and Cycle, model &&model&i";
RUN;
%END;
QUIT;
TITLE;
SYMBOL;
%END;
%finmacro:
%MEND; |
20/11/2012, 17h55
Séries chronologiques : décomposition de Beveridge Nelson
00
Stéphane Colas
lorsque votre problème est réglé !
